This fourth model is a new variation. It still compute the mass-action on the numerator, but attempts to make the flux saturate with the part on the denominator with the part of \(1 + {...}\), where we use handling time and the whole biomass consumption of the predator. \[\begin{align}
F_{ij}^{real} = \frac{F_{ij}^{theoretical}}{1+h_ij * \sum_{i=1} ^{i}{F_{j}^{theoretical}}} && F_{ij}^{theoretical} &= \alpha_{j} * B_i * \frac{B_j}{M_j} \\
\end{align}\]
où
- \(\alpha\) est un paramètre spécifique à un prédateur (mais le prédateur peut se retrouver dans plusieurs réseaux différents)
- \(B_i\) et \(M_i\) sont la biomasse et la bodymass d’une proie dans un réseau spéficique
- \(B_j\) et \(M_j\) sont la biomasse et la bodymass d’un prédateur spéficique. La biomasse du prédateur est la même dans un réseau, mais peut varier d’un réseau à l’autre, alors que son bodymass sera le même pour chacun des réseaux.
et h_ij est
\[\begin{align} h_ij = c \cdot (\frac{Mj}{Mi})^b \end{align}\]
| mean | se_mean | sd | 2.5% | 25% | 50% | 75% | 97.5% | n_eff | Rhat | |
|---|---|---|---|---|---|---|---|---|---|---|
| a_pop | -9.1414577 | 0.0049563 | 0.3302013 | -9.7788607 | -9.3684845 | -9.145091 | -8.9213974 | -8.4840944 | 4438.636 | 1.0005089 |
| a_sd | 2.9889258 | 0.0029802 | 0.2249390 | 2.5933328 | 2.8310624 | 2.973553 | 3.1349945 | 3.4670774 | 5697.049 | 0.9998296 |
| c | -5.0528805 | 0.0039098 | 0.2476439 | -5.5518453 | -5.2191157 | -5.050485 | -4.8834721 | -4.5784168 | 4011.923 | 1.0006896 |
| b | -0.1319591 | 0.0005836 | 0.0398402 | -0.2127973 | -0.1593859 | -0.131059 | -0.1050879 | -0.0550546 | 4660.461 | 1.0005961 |
| sigma | 1.7509523 | 0.0005367 | 0.0433561 | 1.6686601 | 1.7212648 | 1.750266 | 1.7793940 | 1.8407485 | 6526.147 | 0.9997048 |